Eta Carinae
August 2017
Supernova remnants, like Vela Junior, our June Source of the Month are widely believed to be the main accelerators of the ubiquitous cosmic rays that bombard Earth. But what happens just before a massive star is disrupted by a violent supernova explosion at the end of its lifespan?
The possibility to observe this final stage of a massive star's life is rare, but of great importance to understand the last stages of stellar evolution. Moreover, it is relevant to study how massive stars enrich their surroundings with dense gas, which the supernova (or even a gamma-ray burst jet) will plough through and accelerate particles. Luckily, H.E.S.S. had the chance to observe a star in this highly unstable stage, and even detected it for the first time it in very-high-energy gamma rays!
The star we are speaking of is actually two: a binary system, composed of a massive luminous blue variable, with a mass 100 times that of our sun, and a lighter companion star of O- or B-type that is orbiting its primary in a very eccentric orbit. This binary system is observed since at least 4 centuries in more and more observational windows as technology advanced over time.
What captured the attention of our ancestors were drastic changes in luminosity like those that occurred in 1838. Sir John Herschel had been monitoring the system at the Cape of Good Hope from 1834, and did not see any change in luminosity up until 1838, when a sudden increase in brightness occurred [1]. Thomas Maclear and C.P. Smyth, who were also observers at the Cape of Good Hope Observatory, took interest in this variable star and entered into correspondence with Sir Herschel to find all available data. Figure 1 shows some of the original lists of comparisons from 1843 as compiled by Smith & Frew (2011).
Back then called Eta Argus, the binary system is now known as Eta Carinae. It has been the target of numerous observations with ground-based telescopes as well as satellites in radio, millimetre, ultraviolet, infrared, optical and X-ray wavelengths.
Today we know that the period of exceptional brightness that Herschel, Maclear and Smyth noticed in 1838 marked the beginning of the "Great Eruption" [1]. In the Great Eruption, a very large amount of material, corresponding to about 15 times the mass of our sun, was expelled from the primary star [3]. The material, that has been released more than 170 years ago, today forms a structure of gas and dust, which looks like two polar lobes and a collar between them. People dubbed the structure shown in Figure 2 "Homunculus Nebula". The Great Eruption was almost as bright as a supernova, but, in contrast to a supernova, the binary survived. This is why sometimes Eta Carinae is also considered a "supernova impostor" [2].
At the centre of the Homunculus Nebula, the two member stars of Eta Carinae circle each other in a roughly 5.54 year very elliptical orbit. To illustrate the extreme physical scales of this system, Figure 3 shows Eta Carinae compared to our solar system. Put into the solar system, Mercury's orbit would fit into the primary star, whereas the secondary star would be located at the Mars orbit at the time of closest encounter.
When Eta Carinae is observed in soft X-rays (2-10 keV photon energies), the emission changes along the orbit and shows a similar behaviour every 5.54 year orbit. Figure 4 shows the X-ray lightcurve as recorded by different satellites, starting with observations in 1996 and covering almost four full orbits now. When the two stars are closest, the emitted flux is at its maximum, then steeply falling to a minimum after the stars passage and again steeply rising to its usual level. With the ongoing motion of the stars, the flux is then slowly decreasing, reaching a local minimum when the stars are farthest away, and afterwards increases again when the next encounter is approaching [5]. Going to higher X-ray energies, up to 50 keV, the flux seems to be constant throughout the orbit [6].
But what is causing the behaviour of the observed X-ray emission? In Eta Carinae, the stars do not just orbit each other silently, but they drive strong and dense supersonic stellar winds. The primary star loses almost one entire Sun worth of material every 2000 years in its 500 km/s fast wind. The wind of the companion star is less dense than the primary, but is moving considerably faster [7]. When moving along the orbit, the fast wind of the companion clashes into the wind of the primary star and wind material heats up to 50 million degrees, giving rise to the soft X-ray emission we see [5]. Where the winds clash, they form a colliding wind region which is filled with hot shocked gas. Charged particles can be accelerated in these shocks and reach energies that are several orders of magnitude higher than those produced by the heat of the collisions [8]. Since the process responsible for the emission in Eta Carinae is the collision of its winds, it belongs to the source class of colliding-wind binaries.
In 2009, a list of 205 bright sources detected with the newly launched gamma-ray satellite Fermi were announced. Surprisingly, Eta Carinae was among them and showed strong gamma-ray emission up to GeV energies [9]. This was the first time that such high-energy-emission was observed from a colliding-wind binary (although, from a theoretical point of view, particle acceleration and gamma-ray emission up to high energies had been predicted by several people, e.g. [10]). Interestingly, Eta Carinae was spared out prior to its high-energy gamma-ray detection [14]).
Since gas in astrophysical plasmas is usually not heated to temperatures high enough to radiate in gamma rays, the interesting question now is: where does this high-energy emission come from? This question is subject to active research, and the following scenarios have been proposed:
- Protons accelerated in the shocks interact with nuclei, which gives rise to the production of neutral pions and their subsequent decay into gamma rays with MeV-GeV energies [11].
- Only the high-energy part of the Fermi emission is considered to be of hadronic origin, and the bulk of the lower energy gamma rays originates from stellar photons that are up-scattered by ultra-relativistic electrons to gamma-ray energies [8].
Motivated by hints of photon-photon absorption modulating the spectrum of a gamma-ray binary with a compact companion, LS 5039, it was argued that this could also be the case here, with a black-body photon field acting as an absorber to the photon flux [12].
Since then, Fermi continued to observe Eta Carinae and has now covered more than one orbit. Interestingly, the observed gamma-ray emission is variable and extends up to 100 GeV energies and possibly beyond. Unfortunately, the LAT instrument aboard Fermi has a too small detector area to collect more photons at energies beyond 100 GeV.
This is where H.E.S.S. comes into play. The energy range where H.E.S.S. is sensitive overlaps at the lowest energies with the Fermi satellite's range and extends to tens of TeV energies. The detection based on Fermi data was especially exciting for H.E.S.S., since Eta Carinae is only visible in the Southern Hemisphere, where H.E.S.S. is the only operating gamma-ray observatory. If Eta Carinae is accelerating particles up to hundreds of GeV energies, then maybe it can go even higher up and produce gamma-rays that H.E.S.S. could see?
H.E.S.S. looked at Eta Carinae for more than 30 hours between 2004 and 2010 and published the results in 2012 [13]. The first try was sobering. No significant emission could be detected, and it seemed that the energy threshold of the HESS-I telescopes was slightly too high to catch the high-energy end of the Fermi observations. Two things complicated the original H.E.S.S. measurement and increased the energy threshold. Eta Carinae is located far south on the sky, which means that air showers have to be brighter and from higher-energy primary particles to travel through the atmosphere and reach the telescopes. Moreover, as beautiful as the optical images of Eta Carinae and the Carina Nebula are, their optical and UV light adds to the night-sky-background noise in Cherenkov shower images, which further increases the instruments energy threshold.
When the original H.E.S.S. array was expanded by a fifth much larger telescope in 2012, it was clear that the array's energy threshold could be significantly lowered to finally reach into the domain of the Eta Carinae gamma-ray emission. During the next close passage of the two stars in summer 2014, the H.E.S.S. telescopes were on target again, taking about 30 hours of observations in a mode that made sure that the high night-sky background did not damage the sensitive Cherenkov cameras.
Although the instrument did not trigger on night-sky-background photons, some made it into the final data when recording real Cherenkov showers. This poses problems to the analysis of the data and constitutes a background to the measurements that has to be separated from the signal. That is why we extensively studied the impact of the starlight on our analysis products, step by step, starting at the lowest data processing level. And this time we should not get disappointed as is clear from Figure 5!
We are happy to state that we have detected Eta Carinae in very-high energies with a significance of 13.6 sigma. Our data set covers the time when the companion star is approaching the primary star, and a few months around the periastron passage when the stars are as close to each other as depicted in Figure 3. Eta Carinae is detected significantly at each of the two phases. With this detection, the new source class of colliding wind binaries is entering the field of very-high-energy gamma-ray astronomy.
References:
[1] Frew D. J., 2014, The Journal of Astronomical
Data, 10
[2] Smith N., Frew D. J., A Revised Historical Light
Curve of Eta Carinae and Timing of Close Periastron Encounters, 2011,
MNRAS, 415, Fig No 1
[3] Smith N., Ferland G. J., 2007, ApJ, 655, 2
[4] Abraham Z., Falceta-Gonçalves D., Beaklini
P. P. B., 2014, ApJ, 791, 2
[5] Corcoran, M. F., Hamaguchi, K., Liburd, J., et
al. 2015a, in American Astronomical Society Meeting Abstracts,
Vol. 225, American Astronomical Society Meeting Abstracts, 344.15
[6] Sekiguchi A., Tsujimoto M., Kitamoto S., Ishida
M., Hamaguchi K., Mori H., Tsuboi Y., 2009, PASJ, 61, 629
[7] Parkin, E. R., Pittard, J. M., Corcoran, M. F.,
Hamaguchi, K., & Stevens, I. R., 2009, MNRAS, 394, 1758
[8] Farnier C., Walter R. & Leyder J., 2011, A&A,
526, 57
[9] Abdo A. A. et al. (Fermi-LAT collaboration),
2009, ApJS, 183, 46
[10] Reimer A., Pohl M., Reimer O., 2006, ApJ, 644,
1118
[11] Ohm, S., Zabalza, V., Hinton, J. A., & Parkin, E. R.,
2015, MNRAS, 449, L132
[12] Reitberger, K., Reimer, O., Reimer, A., et
al. 2012, A&A, 544, A98
[13] Abramowski A. et al. (H.E.S.S. Collaboration),
2012, MNRAS, 424, 128
[14] Reimer A. & Reimer O., 2009, AIP Conference
Proceedings, Volume 1112, pp. 43-53